FAQ
General Questions
What can Seedx find?
Seedx extracts:
- Seed phrases: BIP39 mnemonics (12, 15, 18, 21, 24 words)
- Private keys: Cryptocurrency keys in various formats
- API keys: From various services
- Tokens: Authentication and session tokens
What file formats are supported?
Text formats:
- Plain text (.txt)
- PDF documents (.pdf)
- Word documents (.docx)
- Excel spreadsheets (.xlsx)
- Archives (.zip, .rar, .7z)
Image formats (with AI):
- PNG, JPG, JPEG, GIF
- Screenshots and photos
Do I need an API key for AI features?
No, OCR is built-in and included with your subscription. Just enable AI mode in settings.
How does typo correction work?
Seedx uses intelligent matching to correct common seed phrase typos:
- Misspelled BIP39 words
- Character substitutions
- Missing or extra letters
AI Integration
Which AI providers are supported?
Seedx uses Quantium’s built-in OCR API for image processing. No external API keys needed - it’s included with your subscription.
How accurate is AI image extraction?
Very accurate for clear images. Factors affecting accuracy:
- Image quality and resolution
- Text clarity and contrast
- Background noise/clutter
- Handwritten vs typed text
Are there any limits on AI processing?
OCR requests are processed sequentially to prevent API overload.
How much does AI processing cost?
OCR is included with your Seedx subscription - no additional costs.
Can Seedx read handwritten seed phrases?
Yes, with AI integration. Accuracy depends on handwriting legibility.
Performance
How fast is Seedx?
Performance varies by input:
- Text files: Fast processing
- Archives: Depends on extraction time
- AI/Images: Limited by API response time
Can I process large directories?
Yes, Seedx can scan entire directory trees:
- Recursive file discovery
- Archive extraction up to 3 levels
- Progress tracking for all files
Why is processing slow?
Common causes:
- AI processing: API calls have latency
- Large archives: Extraction takes time
- Many small files: File system overhead
- Disk I/O: Use SSD for better performance
How much RAM do I need?
RAM requirements depend on your workload:
- Text processing requires less RAM
- Large archives require more RAM
- Monitor usage and adjust accordingly
Detection
What seed phrase languages are supported?
Multiple BIP39 wordlists:
- English (default)
- Spanish
- French
- Italian
- Portuguese
- Czech
- Japanese
- Korean
- Chinese (Simplified/Traditional)
Are partial seed phrases detected?
Seedx looks for complete valid phrases (12, 15, 18, 21, or 24 words). Partial phrases are not extracted.
What private key formats are detected?
- Hexadecimal (64 characters)
- Base58 (Bitcoin WIF)
- Base58Check
- Various altcoin formats
Can Seedx detect encrypted seeds?
No, Seedx finds plaintext seed phrases. Encrypted or encoded seeds require decryption first.
Output
Where are results saved?
Default output structure:
output_dir/
├── seeds.txt # Found seed phrases
├── private_keys.txt # Cryptocurrency keys
├── other_secrets.txt # API keys & tokens
└── csv/
├── seeds.csv # CSV with source info
├── private_keys.csv # CSV with source info
└── other_secrets.csv # CSV with source info
What’s in the CSV files?
CSV files contain the same data as text files plus metadata:
- Source file path where secret was found
- Secret type/name for categorization
Can I customize output format?
Output format is standardized for compatibility. Use the CSV files for source tracking.
Are duplicates removed?
By default, yes. Each unique seed/key appears once in output.
Troubleshooting
No results found
- Verify input files contain seed phrases
- Check file format is supported
- Enable AI for image-based sources
- Review detection settings
- Test with known sample file
AI extraction failing
- Ensure images are readable
- Check image quality and format
- Try individual images first
- Check network connectivity
“Invalid seed phrase” in results
The phrase was detected but may be:
- Missing words
- Misspelled beyond correction
- Invalid word order
- Non-BIP39 format
Archive extraction errors
- Archive may be corrupted
- Incorrect password
- Unsupported compression
- Insufficient disk space
High false positive rate
- Adjust detection sensitivity
- Enable stricter validation
- Filter results by confidence
- Review source data quality
Security
Is my data safe?
Seedx processes files locally on your machine:
- Text/document processing is fully local
- Images sent to Quantium OCR service for processing
- Results stay on your machine
Should I use this on production wallets?
Seedx is for data analysis and recovery. Never expose active wallet seeds unnecessarily.
How do I securely delete results?
- Use secure delete tools
- Overwrite files multiple times
- Clear system caches
- Consider encrypted storage
Best Practices
Recommended workflow
- Organize sources: Separate images from text files
- Text first: Process text files (faster)
- AI second: Process images only if needed
- Review results: Verify findings manually
- Secure storage: Encrypt sensitive results
Handling sensitive data
- Process on air-gapped machine if possible
- Encrypt results immediately
- Limit access to output files
- Delete source files after processing